mcp-google-tasks

# <img src="./assets/a1-logo.svg" alt="A1" width="40"> Google Tasks MCP

**English** | [Русский](./README.ru.md)

[![npm](https://img.shields.io/npm/v/mcp-google-tasks)](https://www.npmjs.com/package/mcp-google-tasks)
[![Glama](https://glama.ai/mcp/servers/A1-x-Tech/mcp-google-tasks/badges/score.svg)](https://glama.ai/mcp/servers/A1-x-Tech/mcp-google-tasks)
[![CI](https://github.com/A1-x-Tech/mcp-google-tasks/actions/workflows/ci.yml/badge.svg)](https://github.com/A1-x-Tech/mcp-google-tasks/actions/workflows/ci.yml)
[![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](./LICENSE)

**A1 Google Tasks MCP** lets an AI app manage Google Tasks in plain language. Plan a project as a task list, break work into subtasks with due dates and notes, mark things done and keep everything in order.

It uses the Google Tasks API with your Google account. It separates reversible completion from permanent deletion and makes the limits of the Tasks API explicit instead of implying that every to-do feature is possible.

- **21 tools.** List, create and update task lists and tasks, move tasks, manage completion, sweep finished work and reach the raw API when a dedicated tool is missing.
- **Connects from the conversation.** Say "connect Google Tasks": the server walks you through the OAuth client, catches Google's redirect on `127.0.0.1` with PKCE and keeps the tokens itself — no config files, no restart.
- **Completion is reversible, deletion is not.** `complete_task` has an undo (`reopen_task`); the delete tools are separate and marked destructive.
- **Due dates are honest.** The API stores only the calendar date — the server never pretends to schedule a time of day; times belong in the notes.
- **Minimal Google scope.** A single `tasks` scope, without Drive, Calendar or Gmail access.

Start with a read-only question:

> Show me my task lists and everything due before Friday in my default list.

[Connect the server](#quick-start) · [Explore use cases](#what-you-can-ask-it-to-do) · [Open technical documentation](#technical-documentation)

---

## See it work in a minute

> **You:** What's still open in the "Website launch" list, and what's due this week?
>
> **Assistant:** Shows the open tasks with their due dates, notes and subtasks. Nothing changes.
>
> **You:** Add a task "Send the launch announcement" due Friday as a subtask of "Marketing".
>
> **Assistant:** Shows the target list, the parent task and the proposed task, then asks for confirmation before creating it.
>
> **You:** Confirm.
>
> **Assistant:** Creates the subtask with a Friday due date. It does not complete or delete anything unless you ask separately.

## Contents

- [Quick start](#quick-start)
- [What you can ask it to do](#what-you-can-ask-it-to-do)
- [How a task changes](#how-a-task-changes)
- [What can change](#what-can-change)
- [Getting access](#getting-access)
- [Configuration](#configuration)
- [Data, limits and background work](#data-limits-and-background-work)
- [Technical documentation](#technical-documentation)
- [Support](#support)

## Quick start

You need Node.js 20+ and a Google account. Credentials are not required at install time — the server connects from the conversation.

1. Add the server to your AI app.
2. Say "connect Google Tasks": the assistant walks you through [creating the OAuth client and approving access](#getting-access) without editing config files.
3. Ask the read-only question above.

<details open>
<summary><strong>Codex</strong></summary>

<br>

**In the app:** open **Settings → MCP servers**, select **Add server**, choose **STDIO**, enter the command `npx -y mcp-google-tasks@latest` and environment variables `GOOGLE_TASKS_CLIENT_ID`, `GOOGLE_TASKS_CLIENT_SECRET`, `GOOGLE_TASKS_REFRESH_TOKEN`, then select **Save** and **Restart**.

**From the command line:**

```bash
codex mcp add google-tasks \
  -- npx -y mcp-google-tasks@latest
```

```bash
codex mcp list
```

[Codex MCP documentation](https://learn.chatgpt.com/docs/extend/mcp?surface=cli)

</details>

<details>
<summary><strong>Claude Code</strong></summary>

<br>

```bash
claude mcp add \
  --transport stdio --scope user google-tasks \
  -- npx -y mcp-google-tasks@latest
```

```bash
claude mcp list
```

[Claude Code MCP documentation](https://code.claude.com/docs/en/mcp)

</details>

<details>
<summary><strong>Claude Desktop</strong></summary>

<br>

The current official path is **Settings → Extensions**. For a custom desktop extension, open **Advanced settings → Extension Developer → Install Extension…**, select a `.mcpb` file and follow the prompts.

This repository currently publishes an npm stdio package and does not contain a `.mcpb` bundle. For Claude Desktop builds that still support local configuration, use the following JSON stdio configuration as a fallback:

```json
{
  "mcpServers": {
    "google-tasks": {
      "command": "npx",
      "args": ["-y", "mcp-google-tasks@latest"]
    }
  }
}
```

In those builds, save it to `~/Library/Application Support/Claude/claude_desktop_config.json` on macOS or `%APPDATA%\Claude\claude_desktop_config.json` on Windows.

[Claude Desktop MCP documentation](https://support.claude.com/en/articles/10949351-getting-started-with-local-mcp-servers-on-claude-desktop)

</details>

<details>
<summary><strong>Cursor</strong></summary>

<br>

Add this to `~/.cursor/mcp.json` on macOS/Linux or `%USERPROFILE%\.cursor\mcp.json` on Windows:

```json
{
  "mcpServers": {
    "google-tasks": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "mcp-google-tasks@latest"]
    }
  }
}
```

[Cursor MCP documentation](https://cursor.com/docs/mcp)

</details>

<details>
<summary><strong>VS Code</strong></summary>

<br>

Run **MCP: Open User Configuration** and add:

```json
{
  "servers": {
    "google-tasks": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "mcp-google-tasks@latest"]
    }
  }
}
```

Check it with **MCP: List Servers**.

[VS Code MCP documentation](https://code.visualstudio.com/docs/agent-customization/mcp-servers)

</details>

## What you can ask it to do

### See your lists and plan the day

- Show my task lists and the open tasks in each one.
- What's due this week in the default list? Include subtasks.
- What changed since yesterday? Include completed and hidden tasks.

### Create and organize work

- Create a "Website launch" list with tasks for design, content and QA.
- Add a task due Friday with notes, nested under the "Marketing" task.
- Move a task to the top of the list, under another task, or into a different list.

### Track completion

- Mark the review task done — and reopen it if the work comes back.
- Sweep every completed task out of the list view in one go.
- Show the hidden tasks that were swept earlier.

## How a task changes

1. Tasks live in **task lists**. Every account has a built-in default list, addressed by the special id `@default`; `create_tasklist` adds more.
2. A task carries a title, notes and a **date-only** due date — the API discards the time portion, so a time of day belongs in the notes.
3. Ordering and nesting change only through `move_task`; the API's `position` field is read-only.
4. Completing a task is reversible: `reopen_task` brings it back. Deleting a task or a list is permanent, and deleting a list removes every task in it.

Recurring tasks cannot be created through the API. Tasks assigned from Google Docs or Chat can be read but not created or moved to another list, and deleting one also removes it in the originating document. There is no server-side search or sorting — the server lists tasks and the AI client filters them.

## What can change

| Operation | What happens | Confirmation boundary |
|---|---|---|
| Read task lists and tasks | Reads lists, tasks and their filters | No change |
| Create a task list or a task | Adds a list or an open task | Changes Google Tasks |
| Update a task or rename a list | Changes a title, notes or due date | Changes a task |
| Move a task | Changes order, nesting or the containing list | Changes a task |
| Complete or reopen a task | Toggles completion status | Reversible change |
| Clear completed tasks | Sweeps every completed task in a list out of view | Destructive |
| Delete a task or a task list | Removes the task, or the list with everything in it | Destructive |
| Raw API request | Can call API methods without a dedicated tool | Potentially destructive |

The AI client controls confirmation prompts. The server marks reads, writes and destructive tools so the client can distinguish an inspection from a live change.

## Getting access

Google Tasks requires OAuth 2.0; an API key is not enough. There are two ways in, and the first one needs no configuration files.

### Connect from the chat (recommended)

Say "connect Google Tasks" and the assistant runs the flow with you:

1. `setup_instructions` prints the checklist: create or select a Google Cloud project, enable **Google Tasks API**, configure the consent screen and create a **Desktop app** OAuth client.
2. Download that client's JSON ("Download JSON") and give the assistant its **path** — `set_client` stores it owner-only. The secret never goes through the conversation.
3. `start_login` returns a Google consent link. Open it **on this machine** and approve; the code comes back to a one-shot listener on `127.0.0.1` (PKCE), never through the chat.
4. `finish_login` exchanges the code and saves the tokens to `~/.config/mcp-google-tasks/credentials.json` (mode 0600) and verifies them with a real Google Tasks API call — so an API that is still switched off is caught right there.

The tokens are re-read on every call, so the connection works immediately — no restart of the AI app. `auth_status` shows what is connected, `logout` revokes and deletes it.

### Environment variables (CI, unattended installs)

1. Create or select a Google Cloud project and enable **Google Tasks API**.
2. Configure the OAuth consent screen and create a **Desktop ap